Transonic compressors with high leading edge Mach numbers are today state of the art in gas turbine design. This holds true for both aero engines and stationary gas turbines. In the early days of highly loaded compressors, the development started from the ideas about supersonic compressors with very high stage pressure ratios. This historical development and the basic ideas are described. The contributions of H. E. Gallus to this development are specially referred to. On that basis, today’s transonic compressors with reduced loading have been developed. The characteristic physics and design features of recent compressors are discussed with respect to aerodynamics, performance, structure mechanics and production technology. This is also done in view of the ideas of the pioneers in this field. Future technology trends of these compressor types as well as new compressor types are presented in the last part of the paper.

In the present paper results of measurements are discussed that give a survey of the flow through rotor and stator of a supersonic compressor stage. In order to analyze the flow, piezoelectric and semiconductor transducers are used for measuring the unsteady pressure distributions along the casing on the one hand and schlieren photography procedure and stroboscopic technique for flow visualization on the other. The pressure transducers are mounted at the casing and shifted along the test region by different series of measurements. The effects of the unsteady flow to the stator produced by the rotor, on the behavior of the flow in the stator channels were analyzed. The oscillation of the pressure at the entrance region of the stator is reduced in axial direction by the cascades, so that the amplitudes diminish at the exit region of the stator. Nevertheless the frequency, induced by the rotor, can be well recognized in the unsteady pressure distribution downstream of the stator. These measurements were completed by the visualization of the flow through the stator. The optical system is working by a modified coincidence method. The starting procedure of the stage up to design speed was observed using shadow and schlieren method at continuous and stroboscopic illumination. The development of the shock waves, the shock-boundary-layer-interaction, and the influence of the rotor-wakes to the stator-flow are discussed.

Regarding the extremely high pressure ratios of jet-engine compressors for the next decade, increasing interest belongs to the further development of supersonic compressors with supersonic relative flow at rotor inlet and supersonic absolute flow at stator inlet. In the past, different suitable design procedures for these components have been developed and tested successfully. However, there is a lack of information concerning the off-design performance of supersonic compressors. The present paper first systematically shows blading and flow path geometry of different experimentally investigated supersonic axial flow compressors. These investigations refer to combinations of characteristic rotors and stators with fixed and variable geometry. A comparison of these geometric data with the main characteristics of the flow pattern shows that, for the investigated stages, the three-dimensional passage geometry has an essential influence on the off-design performance. On the basis of this information semi-empirical models are established for a numerical description of the flow phenomena with predominant influence, as for example shock-, profile-, and endwall boundary layer losses and rotor-stator interactions. For the determination of the off-design performance, these models are incorporated into a streamline curvature calculation method. The computer model established is able to describe the off-design characteristics of the different investigated supersonic compressor stages in the most important operating range.

In the present paper theoretical and experimental results referring to different supersonic rotors are discussed. The theoretical approach, based on the Euler equations of motion, is valid for transonic rotor flows. The shocks are treated separately on the basis of the Rankine-Hugoniot equations. The measurements were performed by time-averaging and time-dependent techniques. Comparing the theoretical approach with the experimental data, special attention is paid to the structure of detached front waves and strong channel shocks. These discontinuity surfaces proved to be remarkably spatially curved, so that the pressure rise in the shocks is lower than in the twodimensional case. With respect to the channel and the rotor outlet flow the presented theoretical results conform to the experimental data, if three-dimensional viscous effects are not dominant. Nomenclature a= velocity of sound c, u, w = absolute, circumferential, and relative velocity c v = specific heat d( )/ds = differential change in streamline direction d/ = increment in streamline direction (r, z plane) /blade = blade force h rot =rothalpy M rel ,M Z -relative, axial Mach number n (e} = normal (unity) vector n/n 0 = speed ratio P(t)*T (t ) = (total)pressure, (total) temperature p = time dependent pressure Q = arbitrary variable r = radius rd$ = increment in circumferential direction R = gas constant, =c p -c v s = entropy S = finite area x,y = profile coordinates (d)z = (increment in) axial direction a.== Mach angle F = circulation 7, tp, \ = flow angle in r-z; r-rdd plane, SI surface e = blade angle M k Q -central difference in /, k direction K -ratio of specific heats p (/) = (total) density a = three-dimensional shock angle co = angular velocity Q =vorticity Subscripts abs, rel u,r,z I S1,S2 sh = absolute, relative system = grid point = circumferential, radial, axial = streamline (r, z plane) = S1, S2 surface = shock

A fast and robust calculation method for turbulent shock boundary-layer interaction is presented which enables the design engineer in quickly estimating the boundary-layer behaviour of a transonic compressor profile section. The separated flow in the vicinity of a shock is described in terms of the boundary-layer properties ahead of the shock and the shock strength itself. The method is incorporated in an integral boundary-layer procedure and coupled with the Euler-equations by the equivalent source concept. Calculations for the flow in transonic and supersonic compressor cascades demonstrate the ability of the present method and show good agreement with boundary-layer properties and Mach number distributions obtained from measurements.